Infosys: Responsible AI to Modernise the Energy Sector

On top of reducing greenhouse gas emissions, renewable energy is set to boost global GDP and create millions of new jobs according to the International Renewable Energy Agency. 

The intermittent nature of wind and solar make creates challenges for the energy industry.

Ashiss Kumar Dash, EVP & Global Head Services, Utilities, Resources, Energy & Enterprise Sustainability at Infosys, says that AI could help to generate more reliable wind and solar output.

However, he feels it must be used responsibly.

Ashiss shares his expertise with Energy Digital.

How can AI impact the energy sector?

AI offers incredible possibilities to modernise and improve the energy sector. The energy sector is asset heavy. AI makes these assets intelligent and self-learning. Whether it’s optimising operations, enabling data-informed decisions or predicting equipment failures before they happen, AI offers significant promise in helping energy companies run smarter and more efficiently. AI will open doors for renewables as well, allowing companies to forecast weather patterns better and generate more reliable wind and solar output. On the side of operations, it powers predictive maintenance and reduces downtime for heavy infrastructure. One of the challenges in the rapidly changing energy sector is grid modelling and design; AI is tremendously helpful in this complex area. 

We’re also seeing generative AI come into play, responding to design questions, ensuring safety in the field, automating reports, answering customer inquiries through virtual agents and even helping engineers simulate new designs through digital twin technologies. AI enables energy companies to balance sustainability with performance, while staying ahead in a transforming regulatory and market environment.

What goes into an AI strategy?

I believe there should not be a separate “AI strategy”. AI should be the business strategy in this sector. A strong AI strategy is really about building a scalable, sustainable foundation that works across the entire organisation. At Infosys, we approach this through five key pillars:

• Tech Incubation – Trying out new tools and ideas to solve real-world energy problems.
• Building the Platform and Ecosystem – Making sure the data, infrastructure and integrations are solid.
• Responsible AI – Setting guardrails for fairness, transparency and compliance from the start.
• Focusing on the Right Use Cases – Prioritising projects that offer clear value and ROI. This could mean reducing downtime, streamlining supply chains or cutting costs through automation.
• Reusable Solutions Across Teams – Creating templates and models that can be adapted across all functions – engineering, finance, HR, legal and beyond.

Beyond these pillars, successful AI deployment comes from how well the strategy is executed and embedded into day-to-day operations. That’s why we establish dedicated “strategy and execution pods” to bring together the right mix of tech, domain and business expertise. These teams don’t just plan; they also build, test and scale solutions quickly. And because AI is constantly evolving, we treat the strategy as a living framework—one that can adapt to new regulations, technologies and market shifts. The goal is to ensure AI delivers real, measurable impact while staying aligned with both the company’s values and its long-term vision.

How has Infosys used AI to support the energy sector?

One great example is our work with a major utility company in North America where we deployed AI models to perform on-demand, bottom-up, long-term grid forecasting and enable smart decision making on capital spend and unplanned outages. The first-of-its-kind solution used open-source ML models, a CIM utility model and scalable cloud architecture to provide significant business benefits to the client:

• Improving load forecasting accuracy to 95%.
• Savings of tens of millions (USD) in electrical infrastructure investment by transitioning out from “annual peak load” and reduced unplanned outage to <1% from previous >5-10%. This was primarily because of efficient and proactive grid investments or upgrade in precise grid areas of electrical infrastructure.
• Forecast “At Will” - anytime, anywhere – for any forecast horizon (10 years for DSP, 15 years scenario study, 22 years for 2045 study).
• Faster processing within twenty-four hours for 650K transformers compared to 85 hours for 6,500 assets (i.e. 25 times faster while using 100 times more nodes.)
• High agility - more agile response to emerging regulatory needs.

Similarly, we have more than 60 AI projects driving remarkable transformation across the energy sector.

What can businesses do to ensure ethical and compliant AI deployment?

Responsible AI is absolutely essential in an age where technological advancement has grown in intelligence, use and popularity like never before. Companies need to make sure their AI systems are fair, transparent and compliant with industry standards, values and regulations.

To get it right, businesses need to put strong governance in place to oversee how AI is built and used, making sure it aligns with governing regulations, guardrails and ethical standards. It’s important to embed responsible practices from the start, thinking about fairness, transparency and bias before any model goes live. Regular monitoring helps catch performance drifts or unintended outcomes, and involving cross-functional teams, like legal, HR and compliance, ensures diverse perspectives are part of the process. Just as important is educating employees on how AI works and the risks involved, so they feel empowered to use it responsibly. Infosys views responsible AI as a way to build trust while advancing innovation, performance and efficiency. 

When done right, it becomes a major driver for long-term value in the energy sector and across industries.